Correction, 2026-10-01: the first version of this page said testing ten cover directions costs under $2. At 6 to 21 credits an image it is 60 to 210 credits, about $0.60 to $2.10 at pack rate. It also stated that most podcast discovery now happens through clips on social and that clips are the main discovery channel for most shows; we have no source for either, so both are now phrased as advice.
A podcast is an audio product with three visual jobs, and they have almost nothing in common: cover art that must read at thumbnail size in a list of hundreds, clip graphics that autoplay muted on social, and a set if the show is filmed. Board them separately, and board the cover art first, because it is the only one that decides whether anyone presses play.
1. Cover art, at the size it is actually seen
Design it at 60 pixels and check it at 60 pixels. Podcast covers are browsed as small squares in a list, and covers that work as posters routinely disappear there.
What survives at that size: high contrast, one focal element, and a short title. What dies: taglines, faces at small scale, detailed illustration, and thin type. If the show name is long, the cover carries a shortened version and the full name lives in metadata.
Collect references by screenshotting an actual podcast app browse view rather than by looking at covers individually. The competitive context is the list, not the artwork.
2. Clip graphics, muted and vertical
If the show is promoted through clips on social, the visual system has to work as a 9:16 video with captions and no sound. Board the template: where captions sit, how the speaker is framed, what the waveform or progress element looks like, and how a guest name appears.
Consistency matters more than beauty here. A recognisable clip format is a channel; a beautiful one-off is a post. See how to caption video for silent viewing for the caption layer, which is doing most of the work.
3. The set, if filmed
Light first, as always: two people, no harsh shadows, and something with depth behind them. Then the background, which needs to survive being the same in every episode for a year. Boards for podcast sets usually over-decorate; a set the audience stops noticing in episode three is the right outcome.
What to generate
Generate: cover art concepts, clip template backgrounds, set concepts before you build, and episode-specific graphics. 8frame prices, a credit at $0.01: an image is 6 to 21 credits (Seedream 5 Lite to Nano Banana Pro), so testing ten cover directions is 60 to 210 credits, about $0.60 to $2.10, which is the right way to find a cover that survives thumbnail size.
Do not generate: the hosts or guests, or a set you present as the real one before it exists.
The order that saves work
Cover art first, because it constrains the palette. Clip template second, because it is where volume happens. Set last, because it is the most expensive and should inherit from the two decisions already made rather than lead them.
FAQ
What size should podcast cover art be? Platforms publish current specifications and they change; check the directory's own requirements. The design rule that does not change is that it must read at thumbnail size.
Do podcasts need video? Not to exist, but if you plan to promote the show with clips on social, the clip template is worth boarding even for an audio-only show.
How many cover directions should I test? Three to five, viewed at 60 pixels next to real competitor covers. Testing at full size tells you nothing about the decision people actually make.
Sources
- 8frame image credit prices (Seedream 5 Lite 6, Nano Banana Pro 21) and the $0.01 pack rate: the 8frame tool registry and content fact sheet, checked 2026-10-01
Design the cover at the size people see it, and the clip template for silence.
Try it without signing up: the 8frame canvas opens on a real board. The canvas is free and unlimited; generation is paid from $19/month.